Abstract: The traditional intrusion detection technology in build a statistical model, rule library management and the detection performance aspects have disadvantages and limitations, it impacts of the intrusion detection system application effect. This article proposes and uses Matlab toolbox to establish a neural network based on BP neural network intrusion detection system. Experiment results show that this algorithm improves the intrusion detection system detection accuracy, reduce the rate of false positives and fail, etc, and has certain advantages of convergence. 

Key words: intrusion detection, BP neural network, Matlab